CUAD
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Legal Contract Revision | CUAD | CQ85.82 | 25 | |
| Retrieval | CUAD | Recall91.25 | 13 | |
| Question Answering | CUAD | ANLS0.2498 | 13 | |
| Context Prioritization | CUAD | R@1294 | 12 | |
| Legal text generation | CUAD | ROUGE-L Score55.77 | 10 | |
| Legal Hallucination Detection | CUAD Matched-subset 120 contracts (test) | FAR7.6 | 6 | |
| Corpus poisoning detection | CUAD | Precision25 | 6 | |
| Retrieval-Augmented Generation | Cuad | Faithfulness76 | 5 | |
| Evidence Selection | CUAD | Relative Evidence Count5.17 | 4 | |
| Binary Classification | CUAD 1.0 (test) | Precision25.6 | 4 | |
| Evidence Selection | CUAD | F1 Score33 | 3 | |
| Contract Review | CUAD (test) | F1 Score88.3 | 3 | |
| Header intrusion detection | CUAD | Accuracy61.7 | 2 | |
| Multi-hop QA | CUAD GE (test) | Token F138.58 | 2 | |
| Multi-hop QA | CUAD DS (test) | Token F137.19 | 2 |
Showing 15 of 15 rows